We describe Fair Layered Increase/Decrease with Dynamic Layering (FLID-DL), a new multi-rate congestion control algorithm for layered multicast sessions. FLID-DL generalizes the receiver-driven layered congestion (RLC) control protocol introduced by Vicisano, Rizzo, and Crowcroft, ameliorating the problems associated with large IGMP leave latencies and abrupt rate increases. Like RLC, FLID-DL is a scalable, receiver-driven congestion control mechanism in which receivers add layers at sender-initiated synchronization points and leave layers when they experience congestion. FLID-DL congestion control coexists with TCP ows as well as other FLID-DL sessions and supports general rates on the dierent multicast layers. We demonstrate via simulations that our congestion control scheme exhibits better fairness properties and provides better throughput than previous methods.

This paper presents and evaluates two principles for designing robust, reliable, and efficient collection protocols. These principles allow a protocol to benefit from accurate and agile link estimators by handling the dynamism such estimators introduce to routing tables. The first is datapath validation: a protocol can use data traffic as active topology probes, quickly discovering and fixing routing loops. The second is adaptive beaconing: by extending the Trickle code propagation algorithm to routing control traffic, a protocol sends fewer beacons while simultaneously reducing its route repair latency. We study these mechanisms in an implementation called Collection Tree Protocol (CTP) and evaluate their contributions to its performance. We evaluate CTP on 12 different testbeds ranging in size from 20–310 nodes and comprising 7 hardware platforms, on 6 different link layers, and on interference-free and interference-prone channels. In all cases, CTP delivers> 90 % of packets. Many experiments achieve 99.9%. Compared to standard beaconing, CTP sends 73 % fewer beacons while reducing topology repair latency by 99.8%. Finally, when using low-power link layers, CTP has duty cycles of 3 % while supporting aggregate loads of 30 packets/minute. 1

Abstract. We review theoretical models of individual motility as well as collective dynamics and pattern formation of active particles. We focus on simple models of active dynamics with a particular emphasis on nonlinear and stochastic dynamics of such self-propelled entities in the framework

Anne SABOURIN Mélanges bayésiens de modèles d'extrêmes multivariés, Application à la prédétermination régionale des crues avec données incomplètes. Sous la direction de:

I am grateful to Luc Segoufin who kindly accepted me to be his PhD student. He introduced me to the problem of query enumeration and encouraged me to look for the answers to all the questions that emerged during our collaboration. He was a truly great advisor, always supportive and available for discussions and advice. I would like to thank Arnaud Durand and Patrice Ossona de Mendez for accepting to review this thesis and for their constructive comments on the earlier versions of this manuscript. Many thanks to Nicole Bidoit and Victor Vianu for accepting to be in my jury. Special thanks go to my Parents, who directed me to follow the scientific path and where fully supportive at all times. Thanks to all the members of LSV – it was a pleasure working with You. I would like to thank the awesome people whom I have been working with in the same office on the 4th floor. Without all the nice riddles and the magnificent French lessons my time there wouldn’t be as chouette as it was. Big thanks go to my son Boris, who was kind enough to sleep through the whole nights, allowing his dad to actually write the thesis down.

is selected for implementation with Planet π4. Planet π4 is improved in several aspects for the purpose of analyzing various network implementations. First, its architecture is restructured and cleaned to allow for an easy replacement of the networking component. Second, its core is modified to work in a

Public reporting burden for this collection of Information is estimated to average 1 hour per response, including the time for reviewing instructions, searching existing data sources, gathering and maintaining the data needed, and completing and reviewing the collection of information. Send comments regarding this burden estimate or any other aspect of this

Messaging is a basic service in sensornets. Yet the unique system and application prop-erties of sensornets pose substantial challenges for the messaging design: Firstly, dynamic wireless links, constrained resources, and application diversity challenge the architecture and protocol design of sensornet messaging; Secondly, complex faults and large system scale introduce new challenges to the design of fault-tolerant protocols. The objective of this dissertation is to address the aforementioned challenges of sensornet messaging. Despite the extensive effort in studying sensornet messaging, the lack of a basic un-derstanding of its essential components has been an obstacle for reliable, efcient, and reusable messaging services in sensornets. To address this problem, one task of this dis-sertation is to identify the basic components of sensornet messaging and to study the re-lated algorithmic design issues. More specically, we propose the messaging architecture SMA that consists of three components: trafc-adaptive link estimation and routing (TLR), application-adaptive structuring (AST), and application-adaptive scheduling (ASC). TLR deals with dynamic wireless links as well as the impact of application trafc patterns on

April 29, 2011Für UnsShe had never before seen a rabbit with either a waistcoat-pocket, or a watch to take out of it, and burning with curiosity, she ran across the field after it Lewis CarrollThe following document, written under the supervision of Dr. reviewed by:

We carry out unprecedented experimental measurements of the frictional drag in turbulent soap-film flows over smooth walls. These flows are effectively two-dimensional, and we are able to create soap-film flows with the two types of turbulent spectrum that are theoretically possible in two dimensions: the “enstrophy cascade, ” for which the spectral exponent α = 3, and the “inverse energy cascade, ” for which the spectral exponent α = 5/3. We find that the functional relation between the frictional drag f and the Reynolds number Re depends on the spectral exponent: where α = 3, f ∝ Re−1/2; where α = 5/3, f ∝ Re−1/4. These findings cannot be reconciled with the classic theory of the frictional drag. The classic theory provides no means of distinguishing between one type of turbulent spectrum and another, and cannot account for the existence of a “spectral link ” between the frictional drag and the turbulent spectrum. In view of our experimental results, we conclude that the classic theory must be considered incomplete. In contrast, our findings are consistent with a recently proposed spectral theory of the frictional drag. In this theory the frictional drag of turbulent flows on smooth walls is pre-